#### October 6, 2026 We've had a record-breakingly quiet year in the gulf, but our luck doesn't look like it is going to hold out. In looking at the track, there is an interesting divergence between the predictions of traditional models and what Google DeepMind has to say about a new system in the gulf. We'll see where this goes. ![[IsaiasOct6Compare.png]] The official NWS forecast seemed to split the difference between the models, a great example of when choosing the "mean of a distribution" ends of representing none of the individuals in the ensemble. ![[IsaiasOct6.jpg]] The context is that Google announced what they claim is a major advance in hurricane track prediction. - [WeatherNext: AI model achieves breakthrough in forecasting cycles](https://deepmind.google/blog/weathernext-ai-model-achieves-breakthrough-in-forecasting-cyclones/) We'll see. #### October 7, 2026 Wednesday morning and the track has moved ... toward the prediction of Google's model. ![[IsaiasOct7Ensemble.png]] The comparable graph from Tuesday was not available, but we can see the that Google DeepMind mean moved as well. So it's not that the other models just switched to Google's perspective. Something in the fundmantals had them all shift overnight. ![[IsaiasOct7.png]]